Genetic Interaction Models

Variable expressivity and epistasis go hand in hand when talking about genetic disorders. Knowing what they mean will help you really understand the kind of complications researchers are up against. In this article I’ll illustrate these concepts using a recently published paper on the causes of autism as an example.

One of the genetic risk factors for autism is a small deletion in a region of the genome called 16p12.1 (That’s just an address telling researchers where to look in the genome). This deletion is enriched in individuals with mental retardation, schizophrenia and autism. The thing is, the deletion is considered a risk factor only, because seemingly healthy individuals also have the deletion. More autistic individuals have the deletion than unaffected individuals, but because it is found in unaffected individuals*, it’s clearly not the sole cause of autism. So if this little deletion is not causing autism, than what is?

Remember that every gene interacts with lots of other genes, at the DNA, mRNA and protein level. When we think about genes it’s more like a complex network than one role of single genes. Sometimes, it’s the combination of specific mutations (or alleles) that cause disease.

Take a look at the diagram on the left. What we consider are two pathways, which means groups of interacting genes. The dots represent a gene and connecting lines mean that somehow those genes influence each other’s activity. The single mutation in scenario A is like the deletion at 16p12.1, it results in a mild to negligible effect. The next two scenarios are particularly interesting.

In part B a second mutation occurs affecting a gene in another pathway. The evidence for this model is that secondary mutations in individuals with the 16p12.1 deletion are varied, they’re all over the genome. This additive effect could result in variable expressivity, i.e. the wide variety of characteristics that we call autism or schizophrenia. It’s because different secondary mutations disrupt different secondary pathways, resulting in autism, plus a variety of other characteristics.

Scenario C highlights epistasis: mutations in genes that belong to the same pathway (i.e. network). The combined affect of these mutations, which may be mild or even benign on their own, lead to more severe conditions.

All this points to the difficulty of uncovering the genetic basis of complex diseases. Many different mutation combinations can lead to variable expressivity and epistatic interactions of normally mild mutations can interact in powerful ways. Uncovering how different mutations interact to cause diseases will be a major undertaking of genetics in the years to come.

*One closer inspection it turned out that supposedly unaffected individuals did report higher incidence of learning disabilities, seizures and other symptoms, but much milder than autism, mental retardation or schizophrenia.

Citations and Image:

Veltman, J., & Brunner, H. (2010). Understanding variable expressivity in microdeletion syndromes Nature Genetics, 42 (3), 192-193 DOI: 10.1038/ng0310-192

Girirajan, S., Rosenfeld, J., Cooper, G., Antonacci, F., Siswara, P., Itsara, A., Vives, L., Walsh, T., McCarthy, S., Baker, C., Mefford, H., Kidd, J., Browning, S., Browning, B., Dickel, D., Levy, D., Ballif, B., Platky, K., Farber, D., Gowans, G., Wetherbee, J., Asamoah, A., Weaver, D., Mark, P., Dickerson, J., Garg, B., Ellingwood, S., Smith, R., Banks, V., Smith, W., McDonald, M., Hoo, J., French, B., Hudson, C., Johnson, J., Ozmore, J., Moeschler, J., Surti, U., Escobar, L., El-Khechen, D., Gorski, J., Kussmann, J., Salbert, B., Lacassie, Y., Biser, A., McDonald-McGinn, D., Zackai, E., Deardorff, M., Shaikh, T., Haan, E., Friend, K., Fichera, M., Romano, C., Gécz, J., DeLisi, L., Sebat, J., King, M., Shaffer, L., & Eichler, E. (2010). A recurrent 16p12.1 microdeletion supports a two-hit model for severe developmental delay Nature Genetics, 42 (3), 203-209 DOI: 10.1038/ng.534

We know that lots of genes are involved in cancer progression. For example, you’ve probably read something like “scientists have found the gene for such-and-such cancer” or something similar. What does that really mean? Are there really genes which cause cancer? And why do we have those genes anyways, if that’s all they do?

Generally speaking, there are two broad classes of what we may call cancer genes. One is Tumor Suppressor Genes and the other are called Oncogenes (onco = tumor). It’s important to know that these are just labels scientist use to describe those genes, it doesn’t mean that those genes exist for the sole purpose of causing or preventing cancer.

They way these genes get those labels is by observing what happens when they are mutated (that is, a nucleotide change, most likely causing a decrease in activity). If more tumors result when a gene is mutated, it’s called a tumor suppressor gene, because we can say that when it’s active it “suppresses” tumors. A gene is called an oncogene if it promotes uncontrolled growth (i.e. like a tumor). So scientists have good reason to focused on these genes, because they know they have some role in tumor progression. The question is to understand what those genes are normally doing, and how, and what happens when they stop working.

Anticancer Genes in normal and tumor cells

But there is another class of genes that is also quite exciting: Anticancer genes. These are genes that function normally in their natural environment (i.e. in the cell types where they are usually expressed). However, when these genes are expressed in tumor cells, they cause cell death (see figure). This is called ectopic (i.e. not in your normal place) expression.

This is very exciting because only tumor cells die (anticancer), the normal surrounding tissue is unaffected. That means there is a potentially very specific way of targeting tumor cells for death, without harming normal tissue. If scientists could control the activity of those genes, it would be a very powerful tool against cancer.

Several of those genes are now in preclinical and/or clinical studies. But before the true therapeutic potential of each gene can be understood, they will need further study. For example:

• Does each anticancer gene only target a single type of tumor, or tumors only of a certain age?
• How specific are anticancer genes to killing only tumor cells? Will they kill normal cells at high doses?
• What are the target pathways that anticancer genes affect?

One of the most confusing observations is that in tumor cells, cell death is inhibited. That’s partly why they can continue to grow and grow, and why they’re so hard to kill. So how are molecular pathways leading to cell death different when anticancer genes are involved? What’s happening at the molecular level?

Clearly, there are lots of open questions and the coming years will provide some exciting results into anticancer gene activity and their therapeutic potential. Keep an eye out for this exciting line of research!

Image and Citation:

Grimm, S., & Noteborn, M. (2010). Anticancer genes: inducers of tumour-specific cell death signalling Trends in Molecular Medicine, 16 (2), 88-96 DOI: 10.1016/j.molmed.2009.12.002